Warfarin therapy is frequently associated with adverse drug events. Past efforts to identify potential prevention strategies have focused on using specific indicators or triggers to detect such events. This study examined more than 9000 errors in warfarin use submitted voluntarily through MEDMARX, a database that tracks medication information from participating hospitals. Investigators discovered that inpatient warfarin-related errors occurred most frequently during transcription/documentation and administration, while outpatient errors occurred most frequently during prescribing and dispensing. Using warfarin data as an example, the authors discuss the utility of voluntary error reporting systems and outline the limitations in their use.